Re: [patch] Cache / File container

Hi again Go ahead and commit it, if you think it's the solution (don't wanna be sarcastic about that, i just have no time to check it by myself :) ) chregu On Thu, 30 May 2002, MARTY Fabien wrote: > (Fabien) > > > File container in PEAR_Cache don't have locking system. So, it's > > > possible, that you get corrupted data-entries under bad circumstances. > > > I have run tests with two scripts running in the same moment (one > > > writing, the second one reading). And it's true that you can get > > > corrupted results when fetching the cache. > > > So I have made a patch to implement locking in the file container. > > > Patch is attached. Can I commit it ? > > (Christian) > > not sure, if this is 100% bulletproof, but better than nothing, i assume :) > > I have tested it 10000 times : one script getting the cache whereas another > was updating the cache with no problem. > > (Christian) > > what about deleting? this is not locked right now, right? > > Right but you can't use flock() to do that. And maybe it's not a real > problem. Because, if the file is deleted while you are reading it, maybe you > can continue to read it because the file isn't really destroyed. It must > depend on the OS ??? > > So, in my opinion, maybe the locking system is not perfect (if someone can > find an example with a problem...), but it's better (for me :-)) than the > current one ! > > > > Fabien > > -- nam...christian stocker adr...bremgartnerstr. 66, ch-8003 zurich pho...+41 1 451 6021 www...http://phant.ch/chregu mob...+41 76 561 8860 ema...chregu@phant.ch wor...+41 1 240 5670 gpg...0x5CE1DECB